Photoelectron multiplier tubes (PMT)
To date, the most frequently used sensors for confocal microscopy are photoelectron multiplier tubes (PMT) that offer low-noise electrical signals over a large range of illumination intensities. The main disadvantage of PMTs is the comparably low quantum efficiency of ~30 % – in the best case, for classical photocathodes and not exceeding 45 % for GaAsP layered photocathodes. As the latter introduce very high instability and render these devices susceptible for damage, there is a severe drawback due to the high cost of these devices.
If a photon is absorbed by the photocathode and if the energy of the photon is sufficient to dissociate an electron from the cathode material, this electron is accelerated by an electrical potential towards the positive electrode. In PMTs, the positive electrode is sequenced in a series of dynodes. The typical potential difference at each step is 100 volts. The electrical signal is finally collected at the anode. Upon incidence of an accelerated electron, the dynode releases a number of electrons, i.e., multiplies the absorbed electron. The multiplication factor depends on the potential difference between the dynodes. The total voltage (cathode vs. anode) is tunable up to approx. 1,000 volts. The number of dynodes is typically 6 … 12, and the voltage is divided by the number of dynodes.
An alternative measurement mode to collecting and integrating the analog signal at the anode is photon counting. Here, the signal is analyzed by a comparator circuit. The peaks representing single secondary electrons caused by a single photon are identified and counted. Although PMTs were the first devices to allow photon counting, they lack high-frequency properties, which means they are only suitable for photon counting for very low light levels (few photons per unit time).
Fig. 1: Left: schematic diagram of a PMT. Upon interaction with a photon, a photoelectron is released from the cathode and accelerated by moderate voltage to the first dynode. Here a few secondary electrons are released and accelerated to the next dynode. The signal is finally collected at the anode. Right: schematic diagram of a HyDTM. Upon interaction with a photon, a photoelectron is released from the cathode and accelerated by high voltage to the semiconductor target. Here the kinetic energy is dissipated at once and the charges are additionally amplified by a multiplication layer (avalanche effect). The signal is finally collected from the anode.
Hybrid detectors (HyD™)
The latest developments in sensor technology have brought forth a chimera of vacuum technology (such as PMTs) and silicon technology (such as avalanche photo diodes, APDs). These hybrids [2] were introduced in confocal microscopy by Leica Microsystems, who named them "HyD"s (pron.: highdees). The HyD sensor is equipped with a GaAsP photocathode, but uses only a single acceleration step as opposed to a sequence of dynodes. The applied voltage is approx. 8,000 volts. This design reduces the sensor’s vulnerability and the risk of damage is decreased by some orders of magnitude – although the quantum efficiency is up to 45 %.
The high kinetic energy of the accelerated electrons is dissipated entirely on a silicon target and immediately gives an approx. 1,500-fold amplification. This is not possible with classical dynodes, which at best allow amplification of 3 to 5-fold (and need to be sequenced for that reason). A multiplication layer, which converts the silicon component into an avalanche diode, finally amplifies the signal up to a measurable strength without an external (noise-generating) amplification circuit.
The LightGate
One of the many advantages of a HyD sensor [3] is the high cutoff frequency, which allows the HyD to operate in photon counting mode even at comparably high intensity levels – anyhow at intensities that are typical for standard fluorescent samples in biomedical research and routine. This fact opens the module for gating the detection signal.
To gain from gated detection, the illumination source has to be pulsed. This is an inherent feature of the white light laser (WLL) introduced for confocal microscope systems in 2007 by Leica Microsystems [4]. This source features eight independent emission bandlets which are independently tunable in both color (currently 470 … 670 nm) and intensity. The source is pulsed at 80 MHz – which is a good start for excitation-tunable FLIM measurements as well.
Fig. 2: Effect of light-gated reflection suppression. The left image shows a profile (xz-)section of a fluorescently labeled cell. Light was collected in the emission band and in the excitation band. The reflections at the slide surface and at the coverslip are obvious (strong horizontal lines). Right image: Without changing emission bands, the light gate fully suppresses these reflections.
Gating the detection signal allows signal collection during the fluorescence emission time only between excitation pulses. By excluding the pulses, the background generated by excitation light is efficiently suppressed and independent of any beam splitting or barrier filtering. This LightGate is a new building block in the concept of the "white confocal", referring to a fully tunable and filter-free spectral optical sectioning device.
Fig. 3: Gating strategies. Left: LightGate collects only emission photons after the excitation pulse. This efficiently suppresses the reflected light independently of the wavelength. Right: Gated STED collects only late emitted photons to ensure restriction to long fluorescence lifetime events in the center of the illumination pattern and by implication increasing resolution further.
Gated super-resolution STED
A second area where HyDs are beneficially implemented in gated mode is STED [5] super-resolution imaging. STED is based on illumination of a diffraction limited spot with fluorophore-exciting wavelength and simultaneous illumination of a ring-shaped area with wavelengths that cause de-excitation (stimulated emission). This arrangement renders only a sub-diffraction area in the excited state and consequently allows probing with higher resolution than diffraction-limited.
In essence, the two light qualities compete for excitation and de-excitation. If there is no de-excitation light, as in the center of the superposition of excitation circle and STED ring, the fluorophores emit fluorescence after the characteristic fluorescence lifetime τ, solely depending on the quantum properties of the fluorophore and environmental influences. Outside the center is a second path for the excited state to return: the stimulated emission. Hence the characteristic lifetime is shortened. The shortening depends on the STED intensity and consequently increases with distance from the center (until the maximum intensity is reached, but that is already at the rim of the excited area).
As a result, the characteristic lifetime of fluorescence is radius-dependent in a STED focus with longest lifetimes in the center. By removing early photons, which are most probably emitted from outside the center, the area of observation is further decreased – which is identical with increased resolution [6].
Fig. 5: Effect of gated STED (from left to right). A 76 nm spaced fluorochrome in DNA origami cannot be separated by standard confocal imaging. In cw STED, the resolution is sufficient to separate 76 nm. When collecting the emission only from 0.5 nm on after the excitation pulse the separation is further increased (resolution improved). Later collection, here from 3.0 nm on, yields additional resolution improvement.
